Lead Analysis
Google launches the Gemini agent, a “universal agent for work” that runs on Claude
One agent plans, calls tools across company systems and delivers finished work in Gmail, Docs and dev tools — private preview, no pricing, and the enterprise agent war just turned three-way
Friday, October 9, 2026: Google Cloud unveiled the Gemini agent Thursday at Gemini at Work 2026 — a single work agent that can answer, act, create and code, running on both Gemini and Anthropic’s Claude models, in private preview with no pricing.
Google Cloud announced the Gemini agent on October 8: one prompt box, and the agent plans, uses tools, connects to company systems and hands back finished work inside Gmail, Docs or development tools. Bloomberg frames it as Alphabet ramping competition with OpenAI, Anthropic and Microsoft for corporate customers. It is the third universal-agent pole in a market Anthropic entered with Claude Cowork/Code and OpenAI is defending with ChatGPT Work, and it arrives with a structural twist: Google opened the agent to Anthropic’s Claude models, so an Indian enterprise can run a multi-model agent under one Google control plane.
The governance design is the enterprise-relevant detail. A policy is written once — “agents may not open documents classified Need to Know” — and Gemini applies it to every agent company-wide instead of per-agent. The agent gets its own mailbox, identity, audit trail and spend caps; coworker agents and industry versions for Financial Services and Legal are in preview, with no price and no general-availability date published. Google’s own framing explains why routing matters now: per-token prices have fallen 98% since 2024 while enterprise AI volume exploded, so “running every simple loop through a premium model quickly breaks corporate budgets.”
The timing is pointed. The same day, WSJ-reported OpenRouter data showed OpenAI’s share of developer and enterprise AI spend climbing from under 25% at the start of 2026 to nearly 50% in September, retaking the OpenRouter weekly spend lead after two and a half years as Anthropic slipped from 75-80% to the high 50s. Anthropic’s IPO roadshow is now reported for the week of November 9 — with its public S-1 still missing from EDGAR and a late-October filing deadline looming — so Google is planting its enterprise flag exactly as the two front-runners wage a pre-IPO share war.
For Indian enterprises the relevance is immediate. Google Workspace and Google Cloud are the default stack for a large share of Indian GCCs, BFSI and mid-market operations, so the Gemini agent plus Claude-inside gives buyers a multi-model agent platform with policy and audit controls that map directly onto the compliance lanes India is now building — MeitY promised Thursday an AI regulation consultation paper within a month (safety, deepfakes, human-first), and company-wide agent policy enforcement is precisely the control India’s binding-lanes proposal anticipates. TCS’s Q2 print the same evening — AI revenue above $3 billion, 10% of topline — shows the demand this platform layer will be sold into.
Macro context arrived as a down-day counterpoint: markets crashed Thursday on US-Iran escalation and Strait-of-Hormuz supply fears — Sensex 71,593.24 (−1,045.46, −1.44%), Nifty 22,232 (−371.25, −1.64%), Brent futures +4.2% to $104.41, rupee to 96.88, VIX up ~10% — with roughly ₹10 lakh crore of investor wealth erased. The AI tape and the macro tape have diverged: platform competition is accelerating while oil reprices risk into the same tightening cycle the RBI started last week.
